Correlation of Acetabular Cup Placement Angles Between an Artificial Intelligence-Powered System Using a Smartphone and Human Measurements

Abstract

Background: An automated measurement system for the placement angles of acetabular cup in total hip arthroplasty prostheses was developed utilizing artificial intelligence (AI) algorithms. The AI-powered system enables immediate measurement by capturing an anteroposterior pelvic X-ray through a smartphone camera. Methods: While developing the AI-powered measurement system, we trained AI utilizing 100 labeled anteroposterior pelvic X-rays to recognize the hip joint and 483 labeled anteroposterior pelvic X-rays to identify anatomical landmarks and the acetabular cup. To validate the AI-powered system, we measured the acetabular cup placement angles of 126 unlabeled post-total hip arthroplasty anteroposterior pelvic X-rays with both the AI-powered system and conventional measurement methods and assessed the correlation between the 2 methods. Results: The Pearson’s correlation coefficients for the acetabular cup placement angles measured using the AI-powered system and conventional method were 0.88 (95% confidence interval, 0.84-0.92, P < .001) in inclination angle and 0.76 (95% confidence interval, 0.67-0.83, P < .001) in anteversion angle, respectively. Conclusions: Both inclination and anteversion angles measured using the AI-powered system showed a strong correlation with angles obtained through conventional methods.

Keywords

• Hip
• Arthroplasty
• Artificial intelligence